Statistical approaches were used to describe and explain the presence or absence of active patterned ground features in a mid-latitude alpine area. The relative rarity of these landforms can be problematic for such analyses but this issue was resolved by a novel spatial sampling strategy that considers only terrain within circles drawn around each observed patterned ground feature, which is then subdivided into a grid. This strategy focuses only on fields with potential patterned ground occurrence and can be used for all statistical studies concerning the occurrence of scattered features over extensive areas. These data were examined using factor analyses (multiple correspondence analysis and hierarchical ascendant classification) blended with a complementary bivariate method to associate patterned ground occurrence with eight environmental variables (elevation, exposure, height-distance ratio, drift, glacier influence, vegetation cover, slopewash and lithology). The absence of active patterned ground is associated with low elevation, glacier absence, no slopewash, a low height-distance ratio and talus fans. Zones where active patterned ground features are present are divided into three homogeneous subzones and are mostly associated with glacier influence, a large height-distance ratio and an absence of vegetation. Their location is controlled by the presence of till from different glacial sequences so that the former and current presence of glaciers favours the development of patterned ground features at the landscape scale. Copyright © 2011 John Wiley & Sons, Ltd.